To investigate the use of ultrasound (US)-guided biopsy of thoracic lesions aiming to determine diagnostic success and complication rates and to identify factors that may affect sample adequacy and safety. This was a retrospective study of consecutive percutaneous US-guided biopsy over 10-year period, including 147 procedures in 146 patients (66±7 years, 83 men, 63 women) with lesions located in the lung (67/147), chest wall (54/147), mediastinum (14/147) and pleura (12/147). Overall diagnostic success, yield for benign and malignant diagnoses and diagnostic success according to lesion location, biopsy type (fine-need aspiration [FNA] or core-needle biopsy [CNB]) and number of specimens were calculated. Presence of complications and effect of age, lesion location, biopsy type, and number of specimens were measured. The overall diagnostic success rate was 90.5% and was similar for malignant (90.6%) and benign (87.5%) diagnoses. Specimen adequacy was similar for FNA and CNB (91.2% and 88.9%, p=0.66); number of specimens did not affect yield. Diagnostic success was highest for mediastinal and chest wall lesions (92.9% and 94.4%) and lowest for pleural lesions (75%), albeit not statistically significant (p=0.45). Complications occurred in 4/147(2.7%) cases, exclusively in lung lesions, and were not associated with any covariates analysed. US-guided biopsy is an effective and safe technique for diagnosis of thoracic lesions, with high diagnostic yield and low complication rate. In the presence of an adequate acoustic window, US guidance can be valuable for diagnosis of peripheral lung and mediastinal lesions. Radiologists performing thoracic biopsy should be encouraged to implement or expand the use of US guidance in their practice.